Agent skill

start

Session startup with context recovery. USE WHEN starting session, after /clear, or recovering from autocompact.

View SKILL.md on GitHub Repository
Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/session-context

SKILL.md

/start - Session Startup & Context Recovery

Gather project context and recover from handoffs automatically.

Data Source Priority

1. Pre-loaded (preferred): Check for "Pre-loaded Session Status" in the SessionStart hook output (appears in system-reminder tags at conversation start with additionalContext). If found, parse the JSON code block and use this data directly—no MCP call needed.

2. MCP fallback: If no pre-loaded data exists, call get_session_status with appropriate level.

Look for a section like this in the startup system-reminders:

# Pre-loaded Session Status

The following session context was gathered at startup...

```json
{
  "level": "standard",
  "integrations": {...},
  "recovery": {...},
  "project": {...},
  "beads_triage": {...}
}

The pre-loaded data has the same structure as the MCP response—format and display it the same way.

Quick Start (MCP Fallback)

typescript
// Fresh session - full context
get_session_status({ level: "full" })

// Context recovery mid-session
get_session_status()  // defaults to minimal

// Restore specific handoff
get_session_status({ handoff: "h7k2m" })

Output Levels

Level Use Case Sections
minimal Context recovery project, harness, beads counts
standard Most session starts + beads_triage, agentmail
full Fresh session + claude_mem references

Auto-Recovery

The plugin tracks your work automatically via hooks. If context is lost (autocompact, /clear), recovery options appear:

Session Marker Detection

If you see a marker like <!-- session:abc123 --> in context but lack detailed memory:

  1. This means autocompact occurred mid-session
  2. Immediately call: get_session_status({ handoff: "abc123" })
  3. The MCP returns full working context
  4. Resume work without asking user - just continue

Rolling Checkpoint

If no marker but a recent checkpoint exists (< 24h), the response includes:

json
{
  "recovery": {
    "available": true,
    "id": "abc123",
    "age": "2h"
  }
}

Offer to restore: "Found recent work checkpoint. Resume?"

Display Format

Restored Session

┌─────────────────────────────────────────────────────────────────┐
│  SESSION RESTORED: {id}                                         │
├─────────────────────────────────────────────────────────────────┤
│  Task: {context.task}                                           │
│  Last active: {age}                                             │
├─────────────────────────────────────────────────────────────────┤
│  RESTORED CONTEXT                                               │
│  • Plan: {plan.path} (cached)                                   │
│  • Files: {files.length} modified                               │
│  • Decisions: {userDecisions.length}                            │
│  • Todos: {todos.length} items                                  │
├─────────────────────────────────────────────────────────────────┤
│  FRESH STATE                                                    │
│  • Branch: {branch} ({uncommitted} uncommitted)                 │
│  • Integrations: {detected integrations}                        │
├─────────────────────────────────────────────────────────────────┤
│  Continuing from: "{nextSteps[0]}"                              │
└─────────────────────────────────────────────────────────────────┘

Normal Session Start

┌─────────────────────────────────────────────────────────────────┐
│  SESSION START                                                   │
├─────────────────────────────────────────────────────────────────┤
│  Project: {project.root}                                        │
│  Branch: {branch} ({uncommitted} uncommitted)                   │
├─────────────────────────────────────────────────────────────────┤
│  Integrations: {list available}                                 │
│  Harness: {if present, show status}                            │
│  Beads: {open} open | {actionable} actionable                   │
├─────────────────────────────────────────────────────────────────┤
│  {recommendations based on state}                               │
└─────────────────────────────────────────────────────────────────┘

Session Tracking

After completing significant work, append a session marker to enable auto-recovery:

markdown
<!-- session:{handoff_id} -->

This marker survives autocompact summarization and enables seamless recovery.

Integrations

The plugin auto-detects and enhances with:

Integration Detection Enhancement
Git Always Branch, uncommitted, recent commits
Beads .beads/ exists Issue triage, actionable work
Harness .claude-harness/ exists Features, memory, loop state
Agent Mail MCP configured Inbox status, file reservations
Claude-Mem MCP configured Observation ID references

Recommendations Logic

Based on state, recommend next action:

  1. Handoff recovered → Continue from nextSteps[0]
  2. Loop in_progress (harness) → Resume implementation
  3. Feature failing (harness) → Fix failing verification
  4. Issue in_progress (beads) → Continue that issue
  5. Actionable work (beads) → Show top picks
  6. No tracked work → Suggest adding work

Related Commands

  • /handoff - Create explicit checkpoint for session end
  • /session-context:handoff - Same as above (full name)

Recommended Agent Skills

Expand your agent's capabilities with these related and highly-rated skills.

majiayu000/claude-skill-registry

agent-ops-spec

Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-state

Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-spec

Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-testing

Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-testing

Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.

163 31
Explore
majiayu000/claude-skill-registry

agent-ops-state

Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.

163 31
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results